Superconductivity of NbSe3

HAL Id: jpa-00209337

https://hal.archives-ouvertes.fr/jpa-00209337

Submitted on 1 Jan 1981

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Superconductivity of NbSe3

A. Briggs, P. Monceau, M. Nunez-Regueiro, M. Ribault, J. Richard

To cite this version:

A. Briggs, P. Monceau, M. Nunez-Regueiro, M. Ribault, J. Richard. Superconductivity of NbSe3.

Journal de Physique, 1981, 42 (10), pp.1453-1459. �10.1051/jphys:0198100420100145300�. �jpa-

00209337�

Superconductivity ^of NbSe3

A. Briggs, ^P. Monceau, ^M. Nunez-Regueiro, ^{M. Ribault (*)} and J. Richard

Centre de Recherches sur les Très Basses Températures, C.N.R.S., ^{B.P. 166} X, 38042 Grenoble Cedex, ^France

(Reçu ^le 6 février 1981, ^{révisé le} 18 mai, accepté ^le 10 juin 1981)

Résumé.

Nous avons mesuré les transitions résistives supraconductrices ^de NbSe3 ^sous champ magnétique

pour des pressions supérieures ^{à celles} nécessaires à la suppression de l’onde de densité de charge qui apparaît

à 59 K à pression ^ambiante. NbSe3 ^{est alors un} supraconducteur ^massif anisotrope ^{avec un} effet Meissner total.

Les anisotropies ^des champs magnétiques critiques ^{sont bien décrites par un modèle de masses} effectives qui ^sont

en bon accord avec celles déduites des mesures de conductivité ou des mesures d’oscillations Shubnikov-de Haas.

A pression ambiante, ^{nous avons} mesuré des transitions avec résistivité nulle pour des échantillons de NbSe3 compactés ^{mais sans} effet Meissner total. Cette supraconductivité n’est pas massive ^{et est} due vraisemblablement

aux barrières entre les domaines formant le cristal.

Abstract.

We report measurements of the resistive superconducting transitions of NbSe3 ^{as a} function of

magnetic field for sufficient pressures ^to destroy the lower CDW. These indicate that NbSe3 ^{is a bulk} anisotropic superconductor ^{with a} total Meissner effect. The critical field anisotropies ^are well described by ^{an effective mass}

model and the values obtained are in good agreement with those obtained by conductivity ^or Shubnikov-de Haas measurements. At ambient pressure ^we have measured zero-resistivity transitions for compacted samples, ^{but do}

not find a total Meissner effect. Hence this superconductivity ^{is not a} bulk effect but is probably associated with barriers between domains in the sample.

Classification

Physics ^Abstracts

62.50

74.10

74.60

Introduction.

Among ^the family ^of metal-transi- tion trichalcogenides intensively studied these last years, only TaSe3 ^and NbSe3 [1] ^{remain metallic at}

helium temperature [2, 3]. TaSe3 ^{has a} superconducting

transition at 2.2 K [4]. NbSe3 undergoes ^{two inde-} pendent charge-density ^wave transitions with asso-

ciated lattice distortions at 145 K (Tl) ^and

59 K (T2) [2, 5]. ^Two other resistive anomalies have been observed : the resistivity drops ^{between 2.2 K}

and 1.5 K and it is constant down to 0.4 K where it decreases again. ^{But down to 7 mK no} zero-resistivity

has been measured. Magnetization measurements indicate that a very small part ^{of the} NbSe3 sample

becomes diamagnetic ^{below 2.2 K where the} resistivity drops [3]. ^A filamentary ^{model of} superconductivity

has been proposed ^to explain ^{this low} temperature behaviour [6]. Under pressure the CDW transition temperatures ^{decrease. It has been shown that}

dT 2/dP = - ^6.25 K/kbar [7]. ^{For this} pressure and above NbSe3 ^{is a bulk} superconductor ^{with a} complete

Meissner effect. Zero-resistivity has been measured

on doped NbSe3 samples ^with impurities ^{of tanta-}

lum [8], ^titanium [9] ^{or zirconium} [10], ^for tempe- (*) Permanent address : Laboratoire de Physique ^des Solides, Université Paris-Sud, ^Bât. 510, ⁹¹⁴⁰⁵ Orsay, ^France.

LE

JOURNAL DE PHYSIQUE

T. 42, ^N° 10, ^{OCTOBRE 1981}

ratures around 1.5 K but no magnetization ^{on these} doped samples ^have yet ^been reported. ^{We have}

observed that a zero-resistivity transition occurs in

compacted pure NbSe3 samples. ^{The critical} tempe-

rature is around 0.9 K.

Hereafter we first review the resistive and magnetic

measurements on pure NbSe3 filaments. Then we

describe the experiments ^on compacted NbSe3 samples

with different size grains. Finally ^we report ^measure-

ments on some NbSe3 monocrystals ^under pressure.

We have determined the critical magnetic ^fields parallel ^and perpendicular ^{to the chain direction.}

NbSe3 under pressure is ^an anisotropic super- conductor which near Tc ^can be described with an

effective mass model. The effective masses deduced from the critical magnetic fields for the different

crystallographic orientations are in good agreement with those obtained by conductivity and Shubnikov- de Haas measurements.

1. Pure NbSe3.

^The drop ⁱⁿ resistivity ^below

2.2 K is sample dependent [3]. ^This drop ^{can be a}

few percent ^{of the} resistivity ^{above 2.5 K but we have}

measured 80 % ^{on a} bundle of filaments. Fleming

does not observe such a drop ^on very pure samples

with resistance ratios of 200 [11]. ^{Below 2.2 K the}

94

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:0198100420100145300

1454

resistivity ^is strongly dependent ^{on the current} density

and the critical current decreases exponentially ^with

temperature. The transition can be suppressed ^with

current densities higher ^{than 1} A/cm2. ^Previous

d.c. initial susceptibility measurements on a NbSe3 sample ^{of 300} mg ^at 50 mK in a magnetic ^{field of}

1 Oe indicated that less than 10- 3 of the volume could be superconducting [3]. ^{More accurate} magne- tization measurements using SQUID magnetometers have been recently reported [12]. ^On NbSe3 samples loosely assembled Buhrman et al. show that a small

diamagnetic contribution _appears around 2 K where the resistivity drops ^{and that at} the lowest temperature (a ^{few tens of} mK) ^the diamagnetic signal ^{is 1} % ^of

-

V/4 ^{n in} agreement with the limit of our previous

measurements. They ^{described a} model where indi- vidual planes ^become superconducting ^{around 2 K.}

At lower temperature ^the coupling ^between planes

would increase leading ^{to a 2D-3D} superconducting

transition.

The temperature dependence ^{of the} superconducting

critical current and the absence of a total Meissner effect exclude a bulk superconductivity ^{and we} proposed ^a filamentary model for the superconducting

transition at 2.2 K [6]. ^However NbSe3 ^is very different

morphologically ^from, for instance (SN)x. ^This poly-

mer is an assembly of fibers of around 100 A diameter which dominate the superconducting properties [13].

Very recently Fung and Steeds [14] ^{have observed}

that the CDW lattice consisted of elongated ^domains

with a typical ^{size of} 2 g ^{x 200 A x 200 A for a} pure

NbSe3 sample. Taking ^{into account} this observation

we have assumed that the CDW order parameter ^is

zero inside the borders between two domains in which the phase of the CDW is constant [15]. ^{At low} temperature these borders become superconducting

which lead to a multiconnexç superconductivity.

This assumption ^{can account for the} sample dependent

behaviour because ^{of the} different repartition ^of

domains in different samples.

2. Compacted NbSe3.

^In compounds ^like NbSe3

where the _easy path for electrons is along ^the chains,

the effect of impurities consists of breaking chains ; consequently ^the transport properties ^are principally

due to these interrupted ^{strands. The} resistivity

measurements between 300 K and 4.2 K that we have made on samples doped ^with titanium, zirconium,

tantalum or after electron irradiation [16] ^{show a} strong increase of the resistivity ^{at low} temperature.

The same behaviour has been observed with proton irradiation [9]. Another method of ^« dirtying » ^the samples ^{is to use} powders with different grain ^sizes.

We have crushed NbSe3 ^{filaments to obtain} powder.

This powder ^was passed through several sieves with

grain size less than 100 g (sample F6) ^{and with} grain

size between 43 03BC ^and 60 g (sample F8). ^The compac- tion of the powder ^was done under an hydraulic press at 100 bars at room temperature. The sample ^dimen-

sions are typically ^{10 x 1.5 x 1 mm3.} The d.c. resis- tance is measured by ^a classical four lead contacts.

The current density ^is typically ^{2 x 10- 2} A/cm2. ^The

insert of figure ^{1 shows the} resistivity variation bet-

ween 300 K and 4.2 K [17]. ^The resistivity ^variation

can be separated ^{in two} parts : ^a semiconducting-like

variation all the more important ^{as the} grain ^{size is}

smaller and the resistivity variation of the _pure NbSe3.

The resistivities at 4.2 K normalized to the room tem-

perature resistivity ^are respectively ^{2.07 for F6 and}

3.54 for F8. It can be seen that anomalies due to the two CDWs at 145 K and 59 K remain which ensures

that, although ^no X-rays determination have been

performed ^{on the} powder, ^{we measure the real} NbSe3 phase.

Fig. ^1.

^Low temperature resistance of compacted powder ^of NbSe3. ^The grain ^size of sample F6 is less than 100 Il ^{and for} sample

F8 between 40 u ^{and 63} Il. In the insert, the resistance variation normalized to room temperature ^{as a} function of temperature for F6, ^{F8 and a} monocrystal ^of NbSe3.

Around 2 K the resistivity saturates. We have measured the resistivity of F6 and F8 in a 3He refri- gerator. ^The resistivity variation is shown in figure ^1.

The resistivity drops ^{to zero for the two} samples. ^If

we define the critical temperature ^{at the} middle of the

transition, Te ^{is 0.8 K} for F6 and 0.9 K for F8. Initial

susceptibility measurements made on F6 show a small

diamagnetism of the order of 6 % ^{down to 200 mK}

and increase of this diamagnetism around 100 mK up ^to 25 % of the total diamagnetism.

In figure ^{2 we show} the variation of the resistivity

of F6 as a function of magnetic ^{field at different tem-}

peratures. ^{Due to} the random orientation of each

Fig. ^2.

Variation of the resistance of F6 normalized to its room

temperature resistance as a function of magnetic ^{field at different} temperatures. In the insert the normalized resistance at 21 kG

as a function of temperature. The resistance increases monotonically

when the superconducting ^{state is} suppressed by ^the magnetic ^field.

grain ^{and the} anisotropy ^{of the} crystal, ^{the critical}

fields are not very well defined. But at sufficiently high magnetic ^{fields it can be seen that the} magneto- resistance follows the same variation which is the normal state one. We can obtain the resistivity ^{in a}

fixed magnetic ^field (21 kG) ^{where the} sample ^{is in}

the normal state (except ^{for the two lowest} tempera-

tures where the values are obtained by extrapolation).

The insert of figure ^{2 indicates that when the} super-

conducting ^{state is} suppressed (by ^the magnetic field)

the normal resistivity ^increases monotonically ^below

2 K.

Our compacted samples ^are pure NbSe3 ^{but disor-}

dered in the sense that _many barriers have been intro- duced. Although ^{the two CDWs are} présent, ^{a zero-}

resistivity ^{is observed below 0.9 K without a} complete

Meissner effect. To ascertain that zero-resistivity ^was

not a surface property ^{of the} compacted sample ^we

cut F6 longitudinally ^{in three} parts ^{and we remeasured} separately these three parts. ^No difference in behaviour from the original specimen ^{was observed.}

Zero resistivity has also been measured on doped NbSe3 samples. ^The resistivity ^{below 2.5 K is current} dependent ^for samples doped ^{with 0.1} % of titanium [9]

and with zirconium [10] although the effect of current is less important than for pure NbSe3 [3]. ^{A super-} conductivity transition was reported ^{for a} NbSe3

with 5 % of tantalum at 1.5 K [8]. ^{It was} stated that the superconductivity ^arose because the lower CDW

was smeared out as similar to the _pressure effect. If this is true, ^{as in the case} of _pressure, a complete

Meissner effect must be observed. However no magne- tization measurements have yet ^been performed ^on doped samples.

Lee and Rice [18] have calculated that the size of the domains where the phase of the CDW is constant decreases with the amount of impurities. ^{So in the}

doped samples ^{as in our} compacted samples ^we expect that there is a great number of normal barriers around CDW domains and therefore a superconduct- ing percolation path ^{at low} temperature leading ^to

a zero-resistance. However there is not a total Meissner effect.

3. NbSe3 ^under pressure.

We have measured the

resistivity ^{of two} single crystals ^of NbSe3 ^under pres-

sure below 4.2 K. The typical dimensions were

5 x 0.02 x 0.005 mm. Precession photographs ^taken

with filtered MoK radiation ensured ^{that the crystals}

were single crystals ^{with the} plane (b, c) ^{in the} plane

of the ribbon. The crystals ^{were mounted on an araldic}

disc in a beryllium-copper ^chester clamp capable ^of retaining pressures up ^to 11 kbar at 300 K resulting

in about 7.5 kbar at 4.2 K. Pressures were measured at room temperature ^and nitrogen temperature ^with

a pressure cycled manganin resistance placed ^near

the specimen using ^the calibration data of Itske- vich [19]. ^{For a fixed room} temperature pressure, the

manganin gauge gave ^a reproducible nitrogen ^tem- perature resistance. The fluid used was an isopentane- methyl 2-pentane mixture. The samples ^{were those}

used for Shubnikov-de Haas oscillations measure- ments that we reported previously [7]. ^{For such}

measurements only ^two probes ^{on the} crystals ^were

necessary which excluded absolute d.c. resistance measurements.

NbSe3 ^becomes superconducting when the lower CDW transition is totally suppressed by pressure. In the insert of figure ^{3 we show the} pressure dependence

of the CDW transition T2 ^{and the resistive} super-

conducting transition. Above 5.5 kbar, NbSe3 ^is

superconducting ^{with a} sharp transition and Tc

decreases when the pressure is increased. ^In the critical pressure range where T2 ^varies sharply with pressure,

T, ^decreases rapidly ^{and a} superconducting ^transition

cannot be detected below 5 kbar. The continuous variation of the resistance for _pressures in this _range could be due to pressure inhomogeneities. ^More hydrostatic measurements with helium _gas are under- way ^to study ^this possible coexistence between _super-

conductivity ^{and CDW state.}

Figure ^{3 shows} the resistive transitions for pressures of 5.5 kbar and 7.2 kbar. At 7.2 kbar, NbSe3 undergoes

a very sharp superconducting transition at 3.4 K with

a width of 0.1 K. The residual resistance of 4 Q is the contact resistance because of the two probes

measurement which varies between 3.2-4 Q for the different pressures. At 5.5 kbar there is ^a break in the resistance variation at 3.5 K followed by ^{a continuous}

decrease of the resistance down to 1.5 K. But if we

extrapolate ^the variation of the magnetic ^critical

fields at H

0 as we report below, ^{we find a critical} temperature ^{of 3.5 ± 0.1 K.} Unfortunately ^{because of}

the two contacts measurements we were unable to measure at ambient pressure the drop ⁱⁿ resistivity

at 2.2 K.

1456

Fig. ^3.

Resistive transitions of a monocrystal ^of NbSe3 ^for

pressures of 5.5 and 7.2 kbar. The measurements were made with two probes and the residual 3-4 Q is the contact resistance. In the insert the variation of the lower CDW transition and the super-

conducting transition as a function of pressure. When the CDW is suppressed NbSe3 ^{becomes a bulk} superconductor.

Fig. ^4.

Magnetization ^{of a} NbSe3 sample formed with _many fibers at 5.4 kbar and 1.37 K.

We have previously reported ^initial susceptibility

measurements of a specimen consisting ^of many fibers

(specimen ^C weight ^{of 0.56} g in reference [7]) ^under

pressure. The specimen ^{showed a} complete ^Meissner

effect above 5 kbar at 2.5 K. The difference between the superconducting ^critical temperatures ^obtained by

resistive and magnetic measurements is not well understood and we are trying ^to clarify ^this point.

Figure ⁴ shows the magnetization ^of sample ^{C at}

1.37 K for a pressure of 5.4 kbar. This magnetization

variation is typical ^{of a} superconductor ^with strong flux trapping.

For magnetic measurements we need to know the

angular orientations. The sample has the chains

parallel ^{to b and its} large ^{face is} parallel ^to (b, c).

For the transverse orientation (H ¹ b) ^{we define the}

azimutal angle 9 ^{between H and c : 9}

⁰ corresponds

to H parallel ^{to c} and ç

n/2 ^{at H} parallel ^{at a*}

(which is different from a for the monoclinic structure).

The polar angle 0 is defined in the plane (a*, b) ^and

0=0 corresponds ^{to H//b.} The resistive transitions in magnetic ^{fields were} performed ^{in two different} configurations : firstly ^{with 0}

n/2 ^and 9

0 at 7.7 kbar, ^{for one} sample only, ^{in the course of our}

Shubnikov-de Haas oscillations measurements [7] ^and secondly ^{at 5.5 and 7.4} kbar with the pressure cell inserted in an electromagnet capable ^of rotating between 0=0 and 0

n/2 ^with 9

n/2, ^{for the two} samples. We have drawn in figure ⁵ typical ^resistive

transitions as a function of magnetic ^{field at the}

pressures of 5.5 and 7.4 kbar for the orientations ç = n/2, ^{0=0 and} 9

n/2, ⁰

n/2. ^In figure ^{6 we}

have drawn the temperature variation of the critical fields for the three orientations investigated ^{for the}

same sample. ^The extrapolation ^of H 1. (0

n/2, ç

0)

at 7.7 kbar gives T,

^{3.25 K} compared ^to Tc

^{3.4 K}

at 7.4 kbar and 3.5 K at 5.5 kbar. This result confirms the decrease of the critical temperature ^for pressures

greater ^{than those} necessary for suppressing ^{the lower}

Fig. ^5.

^Resistive transitions of a monocrystal ^of NbSe3 ^for

pressures of 5.5 and 7.4 kbar ^{as a} function of magnetic field. The orientation of the magnetic field with the crystallographic ^{axis are}

shown on the insert. _ç is the azimutal angle with H in the plane perpendicular ^{to the chain or b} axis, ^{0 is the} polar angle ^{with H} rotating ^{in the} plane (a, b). The resistive transitions correspond

to 9

n/2, ⁰

n/2 (H ¹ b) ^and ç

n/2, 0

⁰ (H//b).

Fig. ^6.

^{Variation of} the critical magnetic ^{fields as a} function of the temperature for three different orientations. At 7.7 kbar for 0 = n/2, w

0. At 5.5 and 7.2 kbar for (p

n/2, ⁰

n/2 ^{and for}

cp

n/2, ^0=0.

-

CDW. The extrapolation ^{towards zero} of the critical field H1 (0

n/2, ç

n/2) ^and H Il (0

0, ç

n/2) gives ^{the same critical} temperatures ^as those obtained in zero field. We observe a slight upward ^curvature

in the temperature variation of H Il for the two pres-

sures and for Hl (0

n/2, ç

n/2) ^at 5.5 kbar. Such behaviour has already ^been reported in two-dimen- sional dichalcogenides and is characteristic of aniso-

tropic superconductors [20]. ^{In table 1 we} give ^the slopes dH.l,ll/dT ^{near T,,} for the three _pressures.

Critical magnetic fields have also been measured in tantalum doped samples but without studying ^the anisotropy ^{in the} transverse orientation [8]. ^{Thus we}

can only compare thé dH)j/dTB

The critical fields in a multilayered superconducting

material can be calculated using the model of Law-

rence and Doniach of Josephson coupled planes [21].

When H is perpendicular ^{to the} planes, the fluxoids have a circular cross section and the vortex current circulates freely ^{as for a bulk} superconductor. ^{In this}

case, the critical field HC2J.

qJo/2 nçn ^where

qJo

hcl2 e is the flux quantum, ç the coherence

length. ^{When H is} parallel ^{to the} layers ^{the vortex}

current circulates inside a layer ^and tunnels from one

layer ^{to another as a} Josephson ^{current. Near} T, the

coherence length ^is much larger ^{than the} layer sepa- ration and an effective mass model can be used. The

fluxoids are ellipsoidals ^and the critical field is defined

by :

01,

Following ^{the same model the} angular dependence ^of He2 between the parallel ^and perpendicular ^orienta-

tions is [22] :

At lower temperature the coherence length may become lower than the layers separation ^{and the nor-}

mal vortex cores are located between the layers.

HC2U ^largely exceeds the bulk value. For T T*

the layers decouple and the critical field diverges.

This cross over temperature ^{between a 3D-2D beha-} viour has been calculated by ^{Klemm et al.} [23] ^and

Boccara et al. [24]. ^The decoupling ^{between the} layers explains ^the upward ^{curvature of} He 2 ^measured experimentally.

This theory ^of Josephson junctions has been extend- ed to filamentary superconductors by ^{Turkevich and}

Klemm [25]. They ^{consider an} assembly of filaments oriented in the Z direction which forms a lattice with dimensions a and b. When H is applied perpen- dicular to the chains, ^the vortex current follows the chain and tunnels by ^the Josephson effect from ^one chain to another as when H is parallel ^{to the} planes

in a layered superconductor. ^{When H is} parallel ^to

the chains the vortex currents are entirely Josephson

tunnel currents and the system ^{behaves as a} Josephson grid. ^Near T, the coherence lengths ^{are much} larger

than the chains separation a ^{and b and a} generalized

effective mass model can be described. In the _per-

pendicular orientation the conductivity ^{is no} longer isotropic ^{and the} angular ^azimutal dependence ^of H C2 ^{becomes :}

Table 1.

Initial slopes of the critical magnetic fields dH/dT ⁱⁿ kG/K.

NbSe3

NbSe3

1458

mx and rny being the effective masses respectively ⁱⁿ

the x and _y directions.

The polar dependence ^of H e2 ^{when H} is rotated from

being perpendicular ^to parallel ^to the chains is :

At low temperatures ^{the vortex core can} fit between the chains. Below T T ^* the chains decouple leading

to the divergence of the critical magnetic field but T * is not the same when H is parallel ^or perpendicular

to the chains.

Such a model can be applied ^to NbSe3. ^{The structure}

can be seen as infinite planes ^{of chains two atoms}

thick separated by ^{Van der} Waals gaps parallel ^to

the (b, c) plane. ^{In the} transverse orientation (0

n/2)

at 7.4 kbar the extrapolation ^{at T}

^{0 of} Hc, gives

a coherence length ç(O)

^{360 A much} larger ^{than the}

chain separation. Consequently the effective mass

model near T, ^is applicable ^to NbSe3. ^{We find that} Hl ^is greater ^{for qJ}

0 than for _ç

n/2 indicating

a lower effective mass in this direction. The ratio between H eJ. ^{(9 = 0, 0 = n/2) and} Hc ^(qJ

^{n/2, 0}

n/2)

is equal ^{to 2 which} corresponds ^{to an effective mass}

ratio of 4. Anisotropy ^{in the} transverse orientation has been measured by magnetoresistance ^{and Shub-}

nikov-de Haas oscillation studies. A ratio of 3 bet-

ween the cyclotron ^{effective masses in the same}

orientations (ç

0, ⁰

n/2) ^and (ç

n/2, ⁰

n/2)

has been measured [26] which is in a relatively good

agreement with the value obtained from the aniso- tropy ^{of the} superconducting critical fields.

We show in figure ^{7 the} polar dependence ^{of the}

critical field as a function of 0. We have drawn in the

same figure the theoretical variation given by equa- tion (4) ^{with the} transverse critical field equals ^to

1.15 k0e. The best fit is obtained with Gx

^0.1820

which is in excellent agreement with the value of

Fig. ^7.

Angular dependence ^{of the critical} magnetic ^{field at}

7.2 kbar as a function of the polar angle ^{0. The curve} is the theoretical variation (Eq. (4)).

0.1849 obtained by the ratio of the critical fields with 9

n/2, ⁰

n/2 ^and ç

0, ⁰

n/2. This model

gives ^an anisotropy of 30 between the effective masses

when the electrons are travelling parallel ^{to b or} parallel ^{to c.} This result is in good agreement ^with the anisotropy ^of conductivity ^measured by Ong ^and

Brill [27].

The upward ^{curvature of} Hie ^Il indicates the strong anisotropy ^of NbSe3 ^{and some} decoupling ^between

chains at low temperature ⁱⁿ agreement ^{with a}

Josephson coupled ^chains description. ^{It must be}

noted that the analysis of the critical fields under pressure shows that NbSe3 ^is three-dimensional near

T, ^{with some} decoupling between chains at lower temperatures. ^{This is the} opposite ^{of the statement}

of Buhrman et al. [12] ^where they expect that ^individual

planes ^{to become} superconducting around 2 K and these planes ^to couple together ^{at lower} temperatures

to form a 3D superconductor.

4. Conclusions.

We have studied the low tem-

perature properties ^of NbSe3. The behaviour under pressure is completely different from that at ambient pressure when we apply sufficient _pressure to destroy

the lower CDW transition (T2

^{59 K at P}

0).

NbSe3 ^{is a bulk} superconductor ^{with a} total Meissner effect. The total diamagnetism has been observed with initial susceptibility measurements and the

magnetization curves are typical ^{of an} irreversible type ^II superconductor ^with strong ^flux trapping. ^Due

to the morphology ^of NbSe3 ^{with a} large anisotropy

in the orientation perpendicular ^{to the chains, the}

effective mass model of Josephson coupled ^{chains is}

reasonable near T,,. ^{From the} anisotropy of the critical fields we find a ratio of effective ^masses in the plane perpendicular ^{to b of 4 and 30 in the} plane (a*, b)

which is in good agreement with values obtained from Shubnikov-de Haas oscillations and conducti-

vity anisotropies.

At ambient pressure, pure NbSe3 ^{shows a} drop ⁱⁿ resistivity ^{at 2.2 K and a very small} diamagnetism ^is

measured at very low temperature. ^Powdered samples

show a zero resistance below 0.9 K. Doped samples

show also zero resistance but no total Meissner effect has yet ^been reported. ^{We consider} that, ^unlike NbSe3

under _pressure, this superconductivity ^{is not a bulk}

effect but due to barriers between domains in the

sample.

Acknowledgments.

We would like to thank A. Meerschaut, ^P. Molinié, ^{and J.} Rouxel from the Laboratoire de Chimie Minérale de Nantes for _pro-

viding ^{us with the} samples ^{and M.} Papoular ^and

M. Renard for helpful discussions.

References

[1] MEERSCHAUT, ^{A. and} ROUXEL, J., J. Less Common Metals 39

(1975) 197.

[2] CHAUSSY, J., HAEN, P., LASJAUNIAS, ^J. C., MONCEAU, P., WAYSAND, G., WAINTAL, A., MEERSCHAUT, A., ^MOLI-

NIE, ^{P. and} ROUXEL, J., ^{Solid State} Commun. 20 (1976) ^759.

[3] HAEN, P., LAPIERRE, F., MONCEAU, P., NUNEZ-REGUEIRO, ^M.

and RICHARD, J., Solid State Commun. 26 (1978) ^725.

[4] SAMBONGI, T., YAMAMOTO, M., TSUTSUMI, K., SHIOZAKI, Y., YAMAYA, ^{K. and} ABE, Y., ^J. Phys. ^Soc. Japan ⁴² (1977)

1421.

[5] FLEMING, ^R. M., MONCTON, D. E. and MCWHAN, ^D. B., Phys. ^{Rev. B 18} (1978) ^5560.

[6] HAEN, P., MIGNOT, ^J. M., MONCEAU, P. and NUNEZ-RFGUEIRO, M., J. Physique Colloq. ³⁹ (1978) ^C6-703.

[7] BRIGGS, A., MONCEAU, P., NUNEZ-REGUEIRO, M., PEYRARD, J., RIBAULT, M. and RICHARD, J., J. Phys. ^{C 13} (1980) ^2117.

[8] FULLER, W. W., CHAIKIN, ^{P. M. and} ONG, N. P., Solid State Commun. 30 (1979) ^689.

[9] FULLER, W. W., Thesis, University ^of California at Los Angeles (1980), unpublished ^and

FULLER, W. W., CHAIKIN, ^{P. M. and} ONG, N. P., Solid State Commun. 39 (1981) ^547.

[10] NISHIDA, K., SAMBONGI, ^T. and IDO, M., ^J. Phys. ^Soc. Japan 48 (1980) ^331.

[11] FLEMING, R. M., Phys. ^{Rev. B 22} (1980) ^5606.

[12] BUHRMAN, ^R. A., BASTUSCHECK, ^C. M., SCOTT, J. C. and KULICK, J. D., ⁱⁿ Inhomogeneous Superconductors, ^edited by ^{D. W.} Gubser, ^{T. L.} Francavilla, S. A. Wolf and J. R. Leibowitz (New York, American Institute of Phy- sics) ^1980.

[13] ^{AZEVEDO, L.} J., CLARK, ^W. G., DEUTSCHER, G., GREENE, ^R. L., STREET, ^{G. B. and} SUTER, ^L. J., Solid State Commun. 19

(1976) 197.

[14] FUNG, ^{K. K. and} STEEDS, ^J. W., Phys. Rev. Lett. 45 (1980) ^1696.

[15] MONCEAU, P., RICHARD, ^{J. and} RENARD, M., Phys. Rev., in press.

[16] MONCEAU, P., RICHARD, ^{J. and} LAGNIER, ^J. Phys. ^{C 14} (1981) 2995.

[17] NUNEZ-REGUEIRO, M., Thesis, Université de Grenoble (1979), unpublished.

[18] LEE, P. A. and RICE, ^T. M., Phys. Rev. B 19 (1979) ^3970.

[19] ITSKEVICH, ^E. S., Pribary ^Tekhn. Eksp. ⁴ (1962) 148, Cryogenics

4 (1964) ^365.

[20] WOOLLAM, ^J. A., SOMOANO, ^{R. B. and} O’CONNOR, P., Phys.

Rev. Lett. 32 (1974) ^712.

[21] LAWRENCE, ^{W. E. and} DONIACH, S., Proceedings of ^{the 12th}

International Conf. ^{on Low} Temperature Physics, ^edited by Eizo Kanda (Academic ^{Press of} Japan, Kyoto) 1971,

p. 361.

[22] MORRIS, R. C., COLEMAN, ^{R. V. and} BHANDARI, R., Phys. ^Rev.

B 5 (1972) ^895.

[23] KLEMM, ^R. A., LUTHER, ^{A. and} BEASLEY, ^M. R., Phys. ^Rev.

B 12 (1975) ^877.

[24] BOCCARA, N., CARTON, ^{J. P. and} SARMA, G., Phys. ^{Lett. 49A} (1974) ^165.

[25] TURKEVICH, ^{L. A. and} KLEMM, R. A., Phys. ^{Rev. B 19} (1979)

2520.

[26] MONCEAU, P., Solid State Commun. 24 (1977) ^331.

[27] ONG, ^{N. P. and} BRILL, ^J. W., Phys. ^{Rev. B 18} (1978) ^5265.